Cloud Simulation Facilities of the Atmospheric Chemistry Group

 

General Facility

This facility was previously an integral component of the CSU Department of Atmospheric Science and remains partially so. Previously known as the Cloud Simulation and Aerosol Laboratory, the building became a multiuse facility during the 1990’s as home for the research and educational laboratories and field program staging areas for the CSU Atmospheric Chemistry group (seehttp://chem.atmos.colostate.edu/AtmChem.htm). Consequently, the facility was renamed for a time as the “Atmospheric Chemistry and Aerosol Laboratory”. The laboratory, pictured in Fig. 1 is located at Christman Field, approximately 1/4 mile from the main facilities of the Atmospheric Science Department on the University's Foothills Campus. The laboratory was originally designed and constructed as a cloud simulation facility for cloud physics and weather modification research. Various parts of the building can be slightly over-pressured to control aerosol contamination. Portions of the laboratory retain a single story with high ceilings (26 ft.) to accomodate the cloud chambers. Upper levels provide additional laboratory and office space. A mechanical shop is also located within the structure. 


Figure 1. The former Cloud Simulation and Aerosol Laboratory (Simlab) and Atmospheric Chemistry and Aerosol Laboratory.

The specialized cloud simulation instrumentation that remain located within the building are a 1 m3 isothermal cloud chamber (now defunct), a 2.0 m3 controlled expansion cloud chamber (dynamic cloud chamber). This basic equipment is supplemented by an 84 ft. high variable flow, vertical dilution tunnel (orange and white stack in Fig. 1) and required specialized instrumentation for sensing temperature, humidity and particle (aerosol, cloud droplet, ice crystal) sizes and concentrations.

 

Isothermal Cloud Chamber Facility

Chamber

The CSU isothermal cloud chamber (ICC) has been described by Grant and Steele (1966), Garvey (1975), and DeMott et al. (1983). It was used primarily for calibrating the potential output of ice nucleating aerosols by cloud seeding generators. The physical chamber and many of its subsystems remain in the laboratory, but are not currently operable. The chamber is a closed cylindrical double shell with an inner shell of 0.64 cm aluminum, 1.37 min diameter and 1.52 m high. The outer shell of 10-gauge steel is 1.68 m in diameter, and the annular space between the shells is filled with polyurethane insulation. Cooling is achieved by a two-phase freon flow in a system of 1.5 cm aluminum tubes welded to the outside of the inner shell at 15 cm spacing. These features were designed to maintain wall temperature within ~ 0.1°C around the chamber walls. The inner shell of the chamber is fitted with a black velvet liner about 15 cm from the walls. Originally installed to minimize frost accumulation and shattering, the liner has proved important in maintaining an experimental volume (~1 m3) with an acceptable cloud density gradient. Fig. 2 shows a chamber schematic and Fig. 3 a photograph.

Figure 2. Schematic diagram of the isothermal cloud chamber. The FSSP and acousticcounter (ice crystal detector) indicated in the diagram were installed temporarily during the early 1980’s for cloud characterization and calibrations.


Figure 3. Photograph of the isothermal cloud chamber, indicating various components and ports. 

Cloud was introduced using the system shown schematically in Fig. 2. Cloud droplets were generated continuously by the atomization of distilled water with an ultrasonic humidifier (Monaghan 670). They were then mixed with cold air and allowed to equilibrate with the chamber while rising through a stand tube in its center. By varying dilution airflow, the liquid water content (LWC) could be varied from 0.3 to 3.0 g m-3 without changing the droplet size appreciably. Temperature within the 1 m3 experimental volume was maintained to within ~ 0.3°C of the “set point” over a range from 0 to -25°C. 

Cloud density was continuously monitored by means of a dewpoint hygrometer. The technique employed was to evaporate a cloud sample and measure its dewpoint temperature; the difference between the saturation mixing ratio corresponding to the dewpoint temperature and that corresponding to the cloud temperature was taken as the liquid water content. Temperatures throughout the system were measured by thermocouples and recorded continuously by a chart recorder or appropriate data logger. Droplet sizes measured using a Particle Measuring Systems Forward Scattering Spectrometer Probe ranged between 6 and 9 mm on all occasions. Representative droplet concentrations were 2100 cm-3 at 0.5 g m-3 and 4300 cm-3 at 1.5 g m-3 LWC. The quasi-steady-state nature of clouds in the ICC allowed nucleation and ice crystal growth to be studied as a function of time. 

Ice crystals settling from the cloud after injection of artificial ice nucleating aerosols were collected on microscope slides and were counted using a Nikon 4/10/20x optical microscope that was isolated in a box cooled by a ¼ hp refrigeration compressor. Bottom and side lamps provided illumination of slides for identifying ice crystals.

Slides were sampled from the chamber periodically until nucleation ceased. Counts were converted to numbers effective per gram of nucleant dispersed (= Yield) using the formula, 

Yield = Nic * (Ac/Av) * (Rd/Rg) * (Ds/Vs) 

Nic: total # ice crystals collected per slide view area 

Ac: chamber cross-sectional area(cm2) 

Av: microscope viewing area(cm2) 

Rd: wind tunnel dilution rate(l min-1) 

Rg: AgI generation rate(g min-1) 

Vs: volume of collected sample (l) 

Ds: sample dilution factor

Access to the chamber is via top and bottom 18” diameter plates. The bottom plate includes the precooler for cloudy air and rotates out underneath the ICC. The top plate lifts off the chamber. A large heater/fan sits on top of the chamber for use over the opened top plate during reconditioning to room temperature after operation.

Ancillary Systems

Refrigeration: The primary refrigeration system for the ICC consists of two Copeland Model EAVA-021A-TAC-800 compressors, a Fisher Controls expansion valve (Series 3560 valve positioner and Type 3561 motion transmitter), and a circulating pump (Corken 2069X). The circulating pump discharges liquid coolant to a “header” reservoir (on top of ICC), and then through a series of 5 expansion valves to feed the coils on the ICC wall. A secondary compressor (Copeland Model ESAM-0033-IAA-001) cools coils in the “precooler” plate in the bottom of the cloud chamber, where cloud enters the system. Refrigeration pressures are monitored using manometer pressure gauges mounted in a control panel adjacent to the cloud chamber. A temperature controller monitors a temperature sensor located in a well in the coolant line. This controller sends a current control condition signal to an electro-pneumatic converter(Leeds and Northrup 10970 series), which then translates the current to a pressure signal for the Fisher Controls expansion valve. A pressure gauge for precooler air flow is on the control panel. It has been calibrated versus total (“cold” air) flow rate and is adjusted with a rotameter. An additional rotameter controls flow of air into the ultrasonic nebulizer (“cloud” air).

Room air cleaning: The ICC room is over-pressured by a large blower, including a large honeycomb HEPA filter, to limit contamination of simulated clouds by ambient and/or generated aerosols. Our experience has been that this provides for room total particle concentrations typically around 200 cm-3.

Dry air system: A compressed dry air system that feeds various laboratory ports provided the 20 to 100 liters per minute (40 lpm is typical) required to force the cloudy/cold air mixture into the chamber and thereby achieve equilibration between the cloudy air and the chamber temperature over the full range of operating conditions. A secondary dessicant cartridge filled with rechargeable alumina-silicate beads was used upstream of the precooler system, in order to assure that the driest air entered this part of the system where temperatures may be as low as -30°C. Dry air was also needed to activate the expansion valve, for cleaning frost/snow from the precooler on a daily basis during operations, and for additional dilution of aerosols.

Vertical Dilution Tunnel: The vertical dilution tunnel permits rapid quenching and dilution of ice nuclei generator effluents, thus simulating field-equivalent generation of aerosols from actual ground-based or airborne ice nucleating aerosol generators. The tunnel outlet is about 25 m above ground, so contamination effects in the immediate vicinity are reduced. The tunnel contains a 60 inch diameter, two stage, axial flow fan rated at 150 Hp. Tunnel flow is 114,200 cfm with maximum fan displacement; natural draft flow without the fan is typically 3530 cfm, but varies with wind speed. The fan exhausts through a converging nozzle to a 45” diameter test section through a flow straightener (honeycomb).Access to the test section is provided just above the flow straightener, by means of two doors fitted with observation ports.Five sampling platforms, spaced at 12’ intervals, are available, giving a effective test section length of 54 feet. Maximum air velocity in the test section is 55 m s-1. Pyrotechnics and airborne (solution combustion) cloud seeding generators can be mounted within the tunnel, below the first sampling platform, or they an be mounted in an enclosure at the tunnel inlet that permits variable flow past the generators to simulate the exact flight speeds of interest. Steady state ground generators are generally operated at the tunnel inlet. Aerosols are colleted downstream of the generator using a variety of rigid and flexible sampling devices, and further dilution with particle-free dry air is sometimes done before aerosols are transported to the cloud chambers. 

Data acquisition 

Data acquisition was never automated. Key temperatures for assuring adequate operational control were monitored using a multi-channel data logger. Data on ice crystal formation was collected manually and recorded on data sheets.

Historical Weather Modification Research in the ICC

The ICC served as a de-facto standard for “calibrating” cloud seeding aerosol generation systems from the late 1960’s until 200. These activities and background on historical development of cloud seeding aerosols are detailed in some of the references listed above. Numerous formal reports detailing specific measurement programs were produced for companies and agencies involved in cloud modification research. Some of these reports are available on request from CSU (in electronic format since about 1995). While the primary product of these calibrations was the Yield value for the particular generator, research at CSU during the 1980’s was notable for the focus on the importance of documenting and understanding the rates of ice crystal formation by ice nucleating aerosols. These studies (e.g., DeMott et al. 1983; Blumenstein et al. 1987; Feng and Finnegan, 1989) used data on the time evolution of ice crystal formation in the ICC to elucidate ice formation mechanisms and engineer ice nuclei to express particular activation characteristics in the atmosphere. 

Dynamic Cloud Chamber Facility

The dynamic cloud chamber and its support systems are shown in the schematics in Figures 4 and 5. Detailed descriptions of this device and examples of experiments performed using it are given in DeMott (1988), DeMott and Rogers (1990), DeMott (1990), DeMott et al. (1990), Jensen-Leute and Kreidenweis (1993), DeMott (1995) and DeMott et al. (1995). Some of the details described in these papers are omitted here. 

The dynamic cloud chamber consists of a 2.0 m3 stainless steel outer pressure vessel which houses a thin cylindrical inner copper liner that is vented to the pressure vessel by small holes in the top and bottom plates. Total experimental working volume is 1.19 m3. In the continuous expansion mode of operation, air is evacuated at a controlled rate from the pressure vessel (using a vacuum pump and a stepping-motor-driven control valve connected to the pressure vessel) to produce expansion cooling of the sample air. The space between the pressure vessel and the inner vessel acts as an expansion plenum which helps to dampen changes in flow rate out of the inner vessel as the pressure control valve cycles. The evacuation rate is controlled by computer, based on pre-determined initial conditions of temperature, pressure, humidity and ascent rate. The simulated ascents are based on equations for dry adiabatic expansion until cloud forms and moist adiabatic expansion thereafter. The program allows for a high degree of flexibility regarding initial chamber conditions and accounts for latent heat release at the lifting condensation level. Computed air parcel coefficients are output to an ascent profile memory system that is achieved (using an air preconditioning system and forced cooling of the inner liner). The temperature of the inner liner is cooled to match the calculated mean air temperature by circulating fluid (LEXSOL, Santa Barbera Chemical Company) through spiral-wound copper tubing that is soldered to the copper liner. Large compressors are used to cool the fluid. A computer-controlled digital valve regulates the rate of introduction of the cooled fluid to the circulating fluid loop. A homogeneous working volume, free from large thermal or vapor gradients is thus made to simulate a wide range of air parcel and cloud conditions. An example of a particular experimental versus programmed continuous ascent profile is displayed in Fig. 6.



Figure 4. Schematic diagram of the CSU dynamic cloud chamber.



Figure 5. Photograph of the CSU dynamic cloud chamber.

The chamber can also be operated in an isothermal/isobaric-mode, or small auxiliary vessels (not shown in Fig. 4) whose internal pressure can be controlled can be used to cause small rapid increases or decreases in chamber pressure when they are vented to the cloud chamber. In this way, humidity excursions of known magnitude (calculated by simple thermodynamics in the closed system) can be made. 

Cloud condensation nuclei (CCN) and ice nuclei for experimentation are generated outside the chamber and can be injected at any point prior to or during an expansion. A small fan inside the chamber is used to induce mixing. Polydisperse or monodisperse CCN particles of various compositions are typically generated from aqueous solution by bubbling filtered air through the solutions. These solution droplets are dried in a diffusion type drier before injection into the cloud chamber. When monodisperse aerosols are desired, the polydisperse sample is input to a differential mobility analyzer (TSI Model 3071) and monodisperse aerosols are extracted. Size classified aerosol particles can be injected into the cloud chamber after the desired initial temperature and humidity conditions have been established. Air filtering is used to reduce particle concentrations inside the chamber to< 1 cm-3 prior to injection of samples. The activity of the ammonium sulfate CCN aerosols commonly used can be predicted theoretically (Fitzgerald, 1975). CCN concentration can be adjusted over the range 10 to 104 cm-3. Total condensation nucleus (CN) concentration is monitored with a CN counter (TSI Models 3010 or 3020). Size distributions and total particle concentrations of aerosols can be obtained by using the classifier in series with the condensation nucleus counter and inverting the data to account for multiply charged aerosol particles (Hagen and Alofs, 1983). 

Various measurement systems are used in experimentation. Temperature is measured continuously using an array of ten copper-Constantan thermocouples (0.508 mm wire) located on the inner liner and four type E fine-wire (12.5 micron) thermocouples are sampled at 10 Hz to measure air temperature 25 cm into the air volume from the inner wall (two each at locations TA1, TA2 in Fig. 4). Pressure is measured with two strain gauge type transducers. Humidity is measured with two optical condensation-type dewpoint hygrometers. A prototype differential absorbtance infrared hygrometer from Ophir Corporation (Nelson, 1982) is sometimes installed in the chamber. It provides a continuous measurement of absolute water vapor concentration, which is used with simultaneous measurements of temperature and pressure to determine relative humidity both below and above water saturation. The useable working ranges and system tolerances of the dynamic cloud chamber are: 

                                                          Pressure                                90kpa to 50kpa +/- 0.05kpa

                                                          Temperature                          30 to -60°C

                                                          relative humidity                     0.1% to >100%

                                                          simulated vertical velocity       0.2 m s-1 to 20 m s-1

Figure 6. Example of program (Tp versus Pressure) versus simulated(T versus pressure) ascent profile in the dynamic chamber. Dewpoint temperature (Td) is also shown.

Cloud droplet sizes and concentrations and their changes in time are measured using a Particle Measuring Systems (PMS) Forward Scattering Spectrometer Probe (FSSP-100). A special sampling system has been designed to draw cloud from the chamber through the laser optics. This sampling system has a number of advantages and avoids some the measurement problems associated with these instruments when they sample the free air stream from aircraft (see, DeMott and Rogers, 1990). Examples of cloud droplet spectra at various times after cloud formation point (using ammonium sulfate aerosols as CCN) are shown in Fig. 7. The FSSP measurements are also used to compute liquid water content. Cloud droplet size spectra and ascent profiles of pressure, temperature, and



Figure 7. FSSP cloud droplet spectra observed at various times after thermodynamic cloud point in Fig. 6.

humidity in slow and rapid expansions can be compared with theoretical predictions of the adiabatic expansion parcel models.The flux of ice crystals from the chamber volume is measured at the base of the chamber with various devices. Most data obtained in the past have utilized an ice particle counter (IPC) developed at CSU. A laser-based device similar to that of Lawson and Stewart (1983) is used for ice crystal detection. A primary difference, however, is that the instrument described by Lawson and Stewart uses transmission/depolarization to detect ice crystals and rolls off to near zero response for particles smaller than about 75 mm. The CSU device was configured instead to detect single particles by extinction in a laser beam. Ice crystals are sampled directly from the base of the cloud volume through a funnel-shaped glass sample tube (10mm o.d. inlet, 0.7 mm o.d. outlet, flow 15 cm3s-1). This air stream crosses a HeNe laser beam (0.7 mm diameter), which falls on a fast response solid state photo detector. The extinction signal should be approximately proportional to the particle cross-section area. Experiments have shown that the technique responds to cloud droplets, ice particles, and electronic noise. A threshold circuit is used to discriminate against both noise and small cloud droplets. Calibration of the ice particle counts was made versus "ground truth" collections onto chilled microscope slides and a video-microscope film loop (see below). 

A PMS 230-X ice crystal probe is also installed in the base of the chamber (not shown in Fig. 4, but visible in Fig. 5). This instrument measures particles in 30 size bins between 10 and 300 mm. This provides the capability to determine ice crystal size distributions. A special sampling system (same as described by Horn, 1984) is used to adapt this aircraft instrument for measurement in the cloud chamber.

A video microscope system (see Fig. 8) supplements the other ice crystal measurements and provides detail on ice crystal habits and morphology. Small, low power, low weight video cameras have recently been used in cloud physics research (Murakami et al., 1987; Tanaka, 1988). Our video microscope system consists of a high resolution CCD video Camera, a remote control zoom lens, a high resolution monitor, an illuminator, and a time lapse video reorder (with date and time encoding). Ice crystals sediment onto 16mm film which moves slowly across the bottom of the chamber liner, then through a hole in the liner, and then past the video microscope which is outside the liner (but inside the pressure vessel). The film forms a loop ~2 m long. The cinema film collection system provides a sample area of up to 14 m2 and the consequent ability to detect crystal concentrations as low as 0.4 per liter with one scan of the film loop. The video microscope system can also be used for sizing crystals and examining their growth habits. 


Figure 8. Video-microscope images of ice crystals in the dynamic cloud chamber.

With any ice particle detector that measures at the base of a sample volume, corrections must be made to directly relate the measured signals to nucleation at a particular temperature during continuous cooling. A slight measurement lag in detecting freshly nucleated ice crystals occurs because crystals must grow and settle to the bottom of the chamber. "Instantaneous" pulse nucleation tests in the chamber using liquid CO2 and dry ice injections into supercooled water clouds have shown a nearly Gaussian response, peaking 30 to 75 s after nucleation, depending on temperature and pressure (which affect crystal growth rate and fall velocity). A deconvolution procedure has been formulated to obtain the true response from the measured ice crystal signal in the chamber. Details are given in DeMott and Rogers (1990) and Rogers and DeMott (1990). 

The current data acquisition system is based around a IBM-compatible 386 computer, and incorporates specially-designed interfaces to handle data from PMS optical probes, the CSU ice particle counter, the infrared hygrometer, state parameters, and other instruments that may be added temporarily or permanently in the future. The software includes user-selectable set-up tables, retainable configuration files and real-time displays of state parameters and hydrometeor spectra (PMS FSSP-100 and 230X probes). Displays are updated one per second and are selectable in real time. Data are recorded to the hard disk.

References

Baumgardner D., J.E. Dye and W.A. Cooper, 1986: The effects of measurement uncertainties on the analysis of cloud particle data. Preprints, Joint Sessions, 23rd Conf. on Radar Met. and Conf. on Cloud Physics, 22-26 Sept., Snowmass, CO, Vol. 3, 313-316. 

Blumenstein, R.R., R.M. Rauber, L.O. Grant and W.G. Finnegan, 1987: Application of ice nucleation kinetics in orographic clouds. J. Clim. Appl. Meteor., 26, 1363-1376. 

DeMott, P.J., 1995: Quantitative descriptions of ice formation mechanisms of silver iodide-type aerosols. Atmospheric Research, 38, 63-99.

DeMott, P.J., A.B. Super, G. Langer, D.C. Rogers, and J.T. McPartland, 1995: Comparative characterizations of the ice nucleus ability of AgI aerosols by three methods. J. Wea. Mod., 27, 1-16.

DeMott, P.J., 1990: Quantifying ice nucleation by silver iodide aerosols. Ph.D. Dissertation, Paper No. 466, Dept. of Atmos. Sci., Colorado State University, Fort Collins, CO, 253 pp.

DeMott, P.J. and D.C. Rogers, 1990: Freezing nucleation rates of dilute solution droplets measured between -30 and -40°C in laboratory simulations of natural clouds. J. Atmos. Sci., 47, 1056-1064. 

DeMott, P.J., D.C. Rogers, and R.P. Lawson, 1990: Improvements to the CSU controlled-expansion cloud chamber. Preprints, AMS Conference on Cloud Physics, 23-27 July, San Francisco, CA., 126-149.

DeMott, P.J., 1988: Comparisons of the behavior of AgI-type ice nucleating aerosols in laboratory-simulated clouds. J. Wea. Mod., 20, 44-50.

DeMott, P.J., W.G. Finnegan and L.O. Grant, 1983: An application ofchemical kinetic theory and methodology to characterize the ice nucleating properties of aerosols used in weather modification. J. Clim. Appl. Meteor., 22, 1190-1203.

Fitzgerald, J.W., 1975: Approximation formulas for the equilibrium size of an aerosol particle as a function of its dry size and composition and the ambient relative humidity. J. Appl. Meteor., 14, 1044-1049. 

Feng, D. and W.G. Finnegan, 1989: An efficient, fast functioning nucleating agent - CompositeAgI-AgCl-NaCl ice nuclei. J. Weather Mod, 21, 41-45. 

Garvey, D.M., 1975: Testing of cloud seeding materials at the cloud simulation and aerosol laboratory, 1971-1973. J. Appl. Meteor., 14, 883-890. 

Grant, L.O. and R.L. Steele, 1966: The calibration of silver iodide generators. Bull. Amer. Meteor. Soc., 47, 713-717. 

Hagen, D.H. and D.J. Alofs, 1983: Linear inversion method to obtain aerosol size distribution from measurements with a differrential mobility analyzer. AerosolSci. and Technol., 2, 465-475.

Horn, R.D., 1984: Electronic circuit modifications for the correction of zero image area scans in the PMS 2D imaging probes. Newsletter on Developments of Airborne Cloud Physics Instruments, June. NCAR, Boulder, CO, 4 pp. 

Jensen-Leute and S.M. Kreidenweis, 1993: Studies of the relationship between submicron marine aerosol and initial marine stratus properties. Atmospheric Science Paper No. 545, Colorado State University, Dept. of Atmospheric Science, Fort Collins, CO, 185 pp.

Lawson, R.P. and R.A. Stewart, 1983: An improved optical ice particle counter. 5th Symp. Meteor. Obs. and Instr., 11-15 April, Toronto, Ont., Canada, 46-53. 

Murakami, M., T. Matsuo, T. Nakayama and T. Tanaka, 1987: Development ofcloud particle video sonde. J. Meteor. Soc. Japan, 65, 803-809. 

Nelson, L.D., 1982:Non-contact sensing of atmospheric temperature, humidity and supersaturation.Preprints, AMS Conf. Cloud Physics,, Chicago, Ill., 293-296. 

Rogers, D.C. and P.J. DeMott, 1990: Measuring Ice Nucleation Rates in a Cloud Chamber -- Correcting for Delays in Crystal Detection. Preprints, AMS Conf. Cloud Physics, , 23-27 July, San Francisco, CA., 146-148.

Squires, P. and P.A. Gillespie, 1952: A cloud droplet sampler for use on aircraft. Quart. J. Roy. Meteor. Soc., 78, 387-393.

Tanaka, T. Matsuo, K. Okada, I. Ichimura, S. Ichikawa and A. Tokuda, 1988: A new airborne device for measurement of cloud particles using video-microscope. 10th Intl. Cloud Physics Conf., Bad Homburg, Ann. Meteor., 25, 341-343.